Solution Architecture · Automation Engineering · API Development · Cloud Deployment · UI Design
Lionlead Real Estate is a Johannesburg-based property firm operating across Sandton and the surrounding suburbs. Like most real estate operations, the firm runs on a combination of deep market knowledge and significant manual administrative effort — and valuation documentation sits at the intersection of both.
Before this system, generating a property valuation document for a client required a skilled agent to spend 20 to 30 minutes pulling a CMA report, selecting comparable properties, applying the firm's internal calculation model, formatting a document in PowerPoint, and sending it manually. The process was repeatable, but it was entirely manual — and every minute spent on it was a minute not spent on revenue-generating activity.
Raiis was brought in to automate that process without disrupting the underlying methodology the firm had built over years of operation.
The challenge was not to replace the firm's valuation approach — it was to codify and automate it. Lionlead's principal had developed a three-input blended valuation formula that combined area-level comparable sales, building-specific recent transaction data, and live market listings to produce a defensible price estimate. The formula worked. The problem was that applying it required manual effort at every stage: sourcing the data, running the calculations, formatting the output, and delivering it.
Additionally, there was no consistency guarantee. Different agents might pull different comparables, apply the formula slightly differently, or produce documents in different formats depending on who was doing the work and how much time they had. For a firm where client trust depends on professional, consistent communication, this was a real risk.
The brief was clear: automate the pipeline from CMA report to client email, preserve the existing formula exactly, and produce a result that every agent could generate in under two minutes.
Raiis designed and built a four-layer automation system.
The first layer is the agent interface — a branded submission form that agents access through a private URL. The form collects the CMA PDF the agent has already downloaded from the CMA Info platform, along with prices and sizes for three to five comparable properties the agent has identified on Property24. The interface is minimal by design: no login, no complex navigation, no training required.
The second layer is the valuation engine — a custom Python-based API deployed on cloud infrastructure. When an agent submits the form, the system reads the CMA PDF and automatically extracts the relevant data: the subject property's details from page one, the comparable area sales table from page three, and the fifteen most recent building or complex sales from page six. It calculates a price per square metre from each data source and averages them, then combines these with the market data the agent entered to produce a final blended valuation. The formula is the same one the firm's principal uses manually — translated into code and applied consistently on every submission.
The third layer is the output generation — a branded PDF document that mirrors the firm's existing valuation template. The document contains the subject property details, both comparable sales tables, the three-input calculation breakdown, and the final estimated price range. It is generated automatically and carries the firm's branding on every page.
The fourth layer is the delivery and logging infrastructure. The generated PDF is emailed directly to the client whose address the agent entered. Simultaneously, the submission is logged to a shared Google Sheet — recording the date, agent name, property, final valuation, and client email — giving the firm's leadership a complete, searchable record of every valuation the system produces.
From form submission to client email received: under two minutes.
The system replaced a 20–30 minute manual process with a two-minute agent workflow. Every valuation now produces the same calculation, the same document structure, and the same professional output — regardless of which agent ran it or how much time they had.
The Google Sheets log gives the firm's leadership visibility into valuation activity that did not previously exist in any structured form. Every submission is timestamped, attributed to an agent, and logged with the calculated result — creating an audit trail the firm can reference and analyse over time.
The system is the first of three planned automation products for Lionlead. Products 02 and 03 — agent placement and lead targeting, and WhatsApp lead qualification — are in active scoping and will build on the infrastructure established in Product 01.
In keeping with our standard approach to client work, we do not disclose the specific tools, libraries, or infrastructure configurations used in production systems. The details above describe what the system does and how it functions from the user's perspective. Questions about technical architecture are welcome directly.
Tell us about your operation. We'll tell you what we'd build.